Llama
Meta's open-weight family; foundation of the OSS ecosystem
Llama is Meta's open-weight family that seeded most of the open LLM ecosystem. Llama 4 (2025) moved to a Mixture-of-Experts architecture (Scout/Maverick/Behemoth) but had a mixed reception, prompting Meta's strategic reset toward Superintelligence Labs. Llama models are the most common base for downstream fine-tuning alongside Qwen.
Maker
Meta AI
License
Open weights
Latest
Llama 4 (MoE)
How it fits the stack
Llama with what it depends on (above) and what it feeds (below). The figure renders as a crawlable diagram and upgrades to an interactive 3D graph as it scrolls into view.
Llama in the AI stack. Llama with its immediate upstream dependencies (top) and downstream dependents (bottom) in the AI value chain. Hover a node in 3D, or read the full relationships below.
Graph data (text) — 8 entities, 8 relationships
- Llama —depends on→ Common Crawl
- Llama —uses→ Common Crawl
- Llama —uses→ FlashAttention
- Llama —hosts→ Hugging Face
- llama.cpp / Ollama / GGUF (edge inference) —uses→ Llama
- On-device / edge inference (NPUs, GGUF) —uses→ Llama
- Perplexity AI —uses→ Llama
- Llama —designs→ Meta AI (FAIR)
Depends on ↑ · 4
Context — capital, rivals, policy · · 1